Supplement of Impact of North America on the aerosol composition in the North Atlantic free troposphere

Transcription

1 Supplement of Atmos. Chem. Phys., 17, , Author(s) 217. This work is distributed under the Creative Commons Attribution 3. License. Supplement of Impact of North America on the aerosol composition in the North Atlantic free troposphere M. Isabel García et al. Correspondence to: Sergio Rodríguez The copyright of individual parts of the supplement might differ from the CC BY 3. License.

2 Expanded uncertainty, % M. I García et al.: Supplement of Impact of North America on the aerosol composition in the North Atlantic FT 1 S1. Uncertainty of the gravimetric method Concentrations of PM x were determined by gravimetry following the EN-1497 procedure (except that filter conditioning was performed at 3 35 % relative humidity instead of 5 %). The combined standard uncertainty (uc), associated to a specific PM x concentration, is expressed as the combination of the individual sources of uncertainty identified in EN-1497; by multiplying uc by the coverage factor k = 2, the expanded uncertainty (U) is obtained. U implies that there is a 95 % probability that the true value lies within ± U of the measured value and it was calculated for individual samples, with the Izaña sampling conditions (sampling time: 8 h; sampler flow: 3 m 3 h 1 ), as represented in Fig. S1. The expanded uncertainty associated for PM x <1 µg m 3 is ± 5 µg m 3. U ( %) will depend on the sample mass, with U>5 % for PM x <1 µg m y = 495 x -1 R 2 = PM x, µg m -3 Fig. S1: Expanded uncertainty (U) of the gravimetric method for individual PM x samples as described in EN The European standard procedure sets the relative humidity (for filter conditioning and during weighing) to 5 % to avoid the effect of water absorption by the filter material, and therefore in the filter mass. The conditioning to 3 35 % used in the long-term program of Izaña is suitable for the low relative humidity of the ambient air and is consistent with the measurements of other aerosol properties (optical properties and number size distribution) that are performed after condition the aerosol sample at relative humidity lower than 4 % according to GAW standardization. masses passed over Africa or came from North America and the Atlantic Ocean. After the classification, we used the output of the Barcelona Supercomputing Centre-DREAM8b atmospheric dust forecast system in order to verify that the westerlies associated back-trajectories were not Saharan dust re-circulated events. From the total of the 41 PM 1 samples, 126 were collected under the westerlies (Fig. S2, 177 under the SAL (Fig. S2 and 98 were Saharan dust re-circulated events (Fig. S2C). Figure S1A shows the backtrajectories used as input for the Median Concentrations At Receptor (MCAR) plots of the PM 1 samples collected under the westerlies influence C) Westerlies SAL Fig. S2: 1 days back-trajectories from 28 to 213 of the chemically analyzed samples collected under ( the westerlies, ( the Saharan Air Layer and (C) Saharan dust re-circulated over the North Atlantic conditions. S2. Westerlies and Saharan air Layer Westerlies (WES) and Saharan Air Layer (SAL) air masses were separated attending to the back-trajectories computed with FLEXTRA. A self-developed software, programmed in MATLAB 7.4 (The Mathworks, Natick, US, scanned the latitude and longitude points for each day classifying samples as WES or SAL. This sorting depended on if the air S3. Meteorological scenario Monthly values of key meteorological fields geopotential heights, winds and omega (at 925, 85 and 7 hpa standard levels) and precipitations rates were determined with the National Centre for Environmental Prediction/National

3 Nov Aug May Jan 2 M. I García et al.: Supplement of Impact of North America on the aerosol composition in the North Atlantic FT Centre for Atmospheric Research (NCEP/NCAR) reanalysis data. Figure S3 shows the averaged (28 213) geopotential height at 85 hpa for the most representative months (January, May, August and November) in order to illustrate the latitudinal shift in the North Atlantic anticyclone C) D) Fig. S3: Geopotential Height at 85 mb averaged from 28 to 213 for ( January, ( May, (C) August and (D) November. The location of Izaña is highlighted (black circle). S4. North America storm season The records of tornado climatology provided by the National Weather Service (National Oceanic and Atmospheric Administration, NOA show that North America is affected by a tornado season, which usually occurs from spring to early summer (March to July). The regional frequency of the tornados is related to the progression of the warm season, moving north from spring to summer ( (Fig. S4. In early spring, tornados affects to Gulf States such as Mississippi and Louisiana and then moves northward affecting Kansas, Nebraska, and the Tennessee Valley region during late spring. Into summer, most of Tornado Alley south central United States is active, and then shifts back southward into the late autumn. Tornados, the most violent of atmospheric convective storms, significantly influence the vertical distribution of aerosols by transporting them from the boundary layer to the troposphere (Barth et al., 215). Ginoux et al. (212) identified the Great Plains located between the east of the Rocky Mountains and the Mississippi River as the largest dust source in United States with a main anthropogenic origin. The maximum monthly average of tornadoes occurrence, from Southern (Texas, Oklahoma, New Mexico) to Central (Kansas, Colorado) plains, is from April to June (Fig. S4 during the study period (28 213) according with the maximum recorded high dust concentrations at Izaña Winter Spring Summer Autumn Region J F M A M J Fig. S4: ( Tornadoes which took place in U.S.A during the study period (28 213). The starting point is represented by colours depending on the season of occurrence. ( Monthly average number of tornadoes during the study period (28 213) in the selected States and the whole Region (TX, NM, OK, KS and CO). TX: Texas; OK: Oklahoma; NM: New Mexico; KS: Kansas; CO: Colorado. Data source: TX NM OK KS CO

4 Ca, OM, EC, =, M. I García et al.: Supplement of Impact of North America on the aerosol composition in the North Atlantic FT th P 5 th P A1) A2) A3) 3 th P nss-so 4 NO 3 -,.4.2 B1) B2) B3).34 C1) C2) C3) NH 4 +, D1) D2) D3) E1) E2) E3) F1) F2) F3).6 G1) G2) G3).3 Al, Sea Salt, 1..5 H1) H2) H3) Fig. S5: Monthly distribution of the percentiles 8, 5 and 3 for the concentrations of ( nss-so = 4, ( NO 3, (C) NH+ 4, (D) EC, (E) OM, (F) Ca, (G) Al and (H) sea salt, from Jan 28 to Aug 213.

5 4 M. I García et al.: Supplement of Impact of North America on the aerosol composition in the North Atlantic FT S5. Percentiles of the inorganic compounds concentration To explain the seasonal evolution of the background (3 th P), median (5 th P) and high (8 th P) concentration events, monthly distribution for the aerosol components concentrations are shown in Fig. S5. The aerosol background (3 th P) composition changes significantly thought the year, although nss-so = 4, NH+ 4 and Al are always part of the total bulk. The remaining chemical species exhibit certain seasonality. Two seasons associated to the latitudinal ranges of the North American outflow and the spatial distribution of the source regions can be well differentiated: (i) Jan-May ( 35 4 N) and Jun-Dec ( 4 55 N). The first season is characterized by an increase of the concentrations of most chemical species (nss-so = 4, NH+ 4, OM, Ca, Al and dust), whereas the second season is mainly associated to an increase in the concentration of EC. Supporting References Barth, M. C., Cantrell, C. A., Brune, W. H., Rutledge, S. A., Crawford, J. H., Huntrieser, H., Carey, L. D., acgorman, D., Weisman, M., Pickering, K. E., Bruning, E., Anderson, B., Apel, E., Biggerstaff, M., Campos, T., CampuzanoJost, P., Cohen, R., Crounse, J., Day, D. A., Diskin, G., Flocke, F., Fried, A., Garland, C., Heikes, B., Honomichi, S., Hornbrook, R., Huey, L. G., Jimenez, J., Lang, T., Lichtenstern, M., Mikoviny, T., Nault, B., OSullivan, D., Pan, L., Peischl, J., Pollack, I., Richter, D., Riemer, D., Ryerson, T., Schlager, H., St. Clair, J., Walega, J., Weibring, P., Weinheimer, A., Wennberg, P., Wisthaler, A., Wooldridge, P., and Zeigler, C.: The Deep Convective clouds and Chemistry (DC3) Field Campaign, B. Am. Meteorol. Soc., 96, , doi:1.1175/bams-d , 215.